Affine motion inheritance with virtual pipeline data unit (vpdu) constraint in video coding

Abstract

An example method includes deriving, responsive to determining to inherit affine motion information for a current block of video data located in a current virtual pipeline data unit (VPDU) from a neighboring block of video data located in a different VPDU, the affine motion information for the current block from two motion vectors of the neighboring block of video data regardless of whether the neighboring block of video data is coded with a 4-parameter affine model or a 6-parameter affine model. The example method further includes determining, based on the derived affine motion information for the current block of video data, a predictor block of video data; and reconstructing the current block of video data based on the predictor block of video data.

What is claimed is: 
     
         1 . A device for coding video data, the device comprising:
 a memory to store the video data; and   one or more processors implemented in circuitry and configured to:
 derive, responsive to determining to inherit affine motion information for a current block of video data located in a current virtual pipeline data unit (VPDU) from a neighboring block of video data located in a different VPDU, the affine motion information for the current block from two motion vectors of the neighboring block of video data regardless of whether the neighboring block of video data is coded with a 4-parameter affine model or a 6-parameter affine model; 
 determine, based on the derived affine motion information for the current block of video data, a predictor block of video data; and 
 reconstruct the current block of video data based on the predictor block of video data. 
   
     
     
         2 . The device of  claim 1 , wherein the affine motion information for the current block of video data comprises control point motion vectors (CPMVs). 
     
     
         3 . The device of  claim 1 , wherein the neighboring block is coded with a 6-parameter affine model represented by three control point motion vectors (CPMVs) and is located in a different VPDU than the current block, and wherein, to derive the affine motion information for the current, the one or more processors are configured to derive a 6-parameter affine motion model of the current block based on two CPMVs of the 6-parameter affine motion model of the neighboring block. 
     
     
         4 . The device of  claim 1 , wherein the neighboring block is coded with a 6-parameter affine model represented by three control point motion vectors (CPMVs) and is located in a different VPDU than the current block, and wherein, to derive the affine motion information for the current, the one or more processors are configured to derive a 4-parameter affine motion model of the current block based on two CPMVs of the 6-parameter affine motion model of the neighboring block. 
     
     
         5 . The device of  claim 1 , wherein the neighboring block is coded with a 6-parameter affine model represented by three control point motion vectors (CPMVs) and is located in a same VPDU as the current block, and wherein, to derive the affine motion information for the current, the one or more processors are configured to derive a 6-parameter affine motion model of the current block based on three CPMVs of the 6-parameter affine motion model of the neighboring block. 
     
     
         6 . The device of  claim 1 , wherein the neighboring block is coded with a 6-parameter affine model represented by three control point motion vectors (CPMVs) and is located in a same VPDU as the current block, and wherein, to derive the affine motion information for the current, the one or more processors are configured to derive a 6-parameter affine motion model of the current block based on three CPMVs of the 6-parameter affine motion model of the neighboring block. 
     
     
         7 . The device of  claim 1 , wherein the device comprises a video decoder. 
     
     
         8 . The device of  claim 1 , wherein the device comprises a video encoder. 
     
     
         9 . The device of  claim 1 , further comprising a display configured to display decoded video data. 
     
     
         10 . The device of  claim 1 , wherein the device comprises one or more of a camera, a computer, a mobile device, a broadcast receiver device, or a set-top box. 
     
     
         11 . A method of coding video data, the method comprising:
 responsive to determining to inherit affine motion information for a current block of video data located in a current virtual pipeline data unit (VPDU) from a neighboring block of video data located in a different VPDU, deriving the affine motion information for the current block from two motion vectors of the neighboring block of video data regardless of whether the neighboring block of video data is coded with a 4-parameter affine model or a 6-parameter affine model;   determining, based on the derived affine motion information for the current block of video data, a predictor block of video data; and   reconstructing the current block of video data based on the predictor block of video data.   
     
     
         12 . The method of  claim 11 , wherein the affine motion information for the current block of video data comprises control point motion vectors (CPMVs). 
     
     
         13 . The method of  claim 11 , wherein the neighboring block is coded with a 6-parameter affine model represented by three control point motion vectors (CPMVs) and is located in a different VPDU than the current block, and wherein deriving the affine motion information for the current comprises deriving a 6-parameter affine motion model of the current block based on two CPMVs of the 6-parameter affine motion model of the neighboring block. 
     
     
         14 . The method of  claim 11 , wherein the neighboring block is coded with a 6-parameter affine model represented by three control point motion vectors (CPMVs) and is located in a different VPDU than the current block, and wherein deriving the affine motion information for the current comprises deriving a 4-parameter affine motion model of the current block based on two CPMVs of the 6-parameter affine motion model of the neighboring block. 
     
     
         15 . The method of  claim 11 , wherein the neighboring block is coded with a 6-parameter affine model represented by three control point motion vectors (CPMVs) and is located in a same VPDU as the current block, and wherein deriving the affine motion information for the current comprises deriving a 6-parameter affine motion model of the current block based on three CPMVs of the 6-parameter affine motion model of the neighboring block. 
     
     
         16 . The method of  claim 11 , wherein the neighboring block is coded with a 6-parameter affine model represented by three control point motion vectors (CPMVs) and is located in a same VPDU as the current block, and wherein deriving the affine motion information for the current comprises deriving a 6-parameter affine motion model of the current block based on three CPMVs of the 6-parameter affine motion model of the neighboring block. 
     
     
         17 . The method of  claim 11 , wherein coding comprises decoding. 
     
     
         18 . The method of  claim 11 , wherein coding comprises encoding. 
     
     
         19 . A computer-readable storage medium having stored thereon instructions that, when executed, cause one or more processors to:
 derive, responsive to determining to inherit affine motion information for a current block of video data located in a current virtual pipeline data unit (VPDU) from a neighboring block of video data located in a different VPDU, the affine motion information for the current block from two motion vectors of the neighboring block of video data regardless of whether the neighboring block of video data is coded with a 4-parameter affine model or a 6-parameter affine model;   determine, based on the derived affine motion information for the current block of video data, a predictor block of video data; and   reconstruct the current block of video data based on the predictor block of video data.   
     
     
         20 . The computer-readable storage medium of  claim 19 , further storing instructions that cause the one or more processors to:
 retain, after coding the different VPDU that includes the neighboring block of video data, only a subset of the affine motion information of the neighboring block of video data.